Image processor having object recognition ability

Claim 2

An image processor having object recognition ability for inspecting, recognizing and determining an object with use of images, comprising: an input image memory (12) for storing input digital image information; a binary process means (13) for processing density data of each pixel in said input image memory (12) with threshold value so as to convert the density data to binary image comprised of active images and non-active images; a labeling process means (14) for providing the same label to a group of active images which is adjacent said binary image; a listing process means (15) for producing a list per each label and coordinate information of all active images which belong to said label; a center of gravity calculation means (16) for obtaining the center of gravity O x ,y of each label by averaging individual coordinate in the group of the active image of each label; a coordinate distance/angle calculation means (17) for calculating a coordinate distance d of a straight line which is disposed between the center of gravity O x ,y and the furthest point P x ,y of the label and an angle θ at the center of gravity O x ,y relative to the straight line; a pair label detection means (18) for detecting whether or not another adjacent label forms a pair with said label by checkinq whether or not there is an active image group which belongs to another label within a predetermined semi-circular range which is defined by a distance less than five times of said coordinate distance d and by said angle θ which is less than +/-5°; and an output memory (20) for storing label numbers which are determined to be paired.

Claim 3

An image processor having object recognition ability which inspects, recognizes and determines an object by using images, comprising: an input image memory (12) for storing input digital image information; a binary process means (13) for processing density data of each pixel in said input image memory (12) with threshold value so as to convert the density data to binary image comprised of active images or non-active images; a labeling process means (14) for providing the same label to a group of active images which is adjacent said binary image; a listing process means (15) for producing a list per each label and coordinate information of all active images which belong to said label; a center of gravity calculation means (16) for obtaining the center of gravity O x ,y of each label by averaging individual coordinate in the group of the active image of each label; a coordinate distance/angle calculation means (17) for calculating a coordinate distance d of a straight line which is disposed between the center of gravity O x ,y and the furthest point P x ,y of the label and an angle θ at the center of gravity O x ,y relative to the straight line; a pair label detection means (18) for detecting whether or not another adjacent label forms a pair with said label by checking whether or not there is an active image group which belongs to another label within a predetermined semi-circular range which is defined by a distance less than five times of said coordinate distance d and by said angle θ which is less than +/-5°; and a broad line production means (19) for drawing a broad line by converting an area adjacent said straight line disposed between each center of gravity of each label to active images; and an output image memory (21) for storing all images forming said broad line.

Claim 4

An image processor having object recognition ability which inspects, recognizes and determines an object by using images, comprising: an input image memory (12) for storing input digital image information; a binary process means (13) for processing density data of each pixel in said input image memory (12) with threshold value so as to convert the density data to binary image comprised of active images or non-active images; a labeling process means (14) for providing the same label to a group of active images which is adjacent said binary image; a listing process means (15) for producing a list per each label and coordinate information of all active images which belong to said label; a center of gravity calculation means (16) for obtaining the center of gravity O x ,y of each label by averaging individual coordinate in the group of the active image of each label; a coordinate distance/angle calculation means (17) for calculating a coordinate distance d of a straight line which is disposed between the center of gravity O x ,y and the furthest point P x ,y of the label and an angle θ at the center of gravity O x ,y relative to the straight line; a pair label detection means (18) for detecting whether or not another adjacent label forms a pair with said label by checking whether or not there is an active image group which belongs to another label within a predetermined semi-circular range which is defined by a distance less than five times of said coordinate distance d and by said angle θ which is less than +/-5°; and a broad line production means (19) for drawing a broad line by converting an area adjacent said straight line disposed between each center of gravity of each label to active images; and an output image memory (21) for storing all images forming said broad line.

Claim 5

An image processor having object recognition ability of claim 2, it is characterized in that said broad line production means (19) forms a broad line containing carriers and their nearby pixels which are converted into active images, wherein said pixels are located along the line connecting each center of gravity of said paired labels.

Claim 6

An image processor having object recognition ability of claim 2, it is characterized in that said broad line production means (19) forms a broad line containing carriers and their nearby pixels which are converted into active images, wherein said pixels forms a line connecting each end of said paired labels in the shortest distance.

Claim 7

An image processor having object recognition ability of claims 1, 3 or 4, it is characterized in that said binary process means (13), labeling process means (14), listing process means (15), center of gravity calculation means (16), coordinate distance/angle calculation means (17) and pair label detection means (18) are structured in a computer (24).